No abstract

Background: Atrial fibrillation (AF) burden and duration appear to be related to stroke risk. A wearable consumer electronic device could provide long-term assessment of these measures inexpensively and noninvasively. This study compares the accuracy of an AF-sensing watch (AFSW; Apple Watch with KardiaBand) with simultaneous recordings from an insertable cardiac monitor (ICM; Reveal LINQ). Methods: SmartRhythm 2.0, a convolutional neural network, was trained on anonymized data of heart rate, activity level, and ECGs from 7500 AliveCor users. The network was validated on data collected in 24 patients with ICMs and a history of paroxysmal AF who simultaneously wore the AFSW with SmartRhythm 0.1 software. The primary outcome was sensitivity of the AFSW for AF episodes ≥1 hour. Secondary end points included sensitivity of the AFSW for detection of AF by subject and sensitivity for total AF duration across all subjects. Subjects with >50% false-positive AF episodes on ICM were excluded. Results: We analyzed 31 348.9 hours (mean (SD), 11.3 (4.4) hours/day) of simultaneous AFSW and ICM recordings in 24 patients. The ICM detected 82 episodes of AF ≥1 hour while the AFSW was worn, with a total duration of 1127.1 hours. Of these, the SmartRhythm 2.0 neural network detected 80 episodes (episode sensitivity, 97.5%) with a total duration of 1101.1 hours (duration sensitivity, 97.7%). Three of the 18 subjects with AF ≥1 hour had AF only when the watch was not being worn (patient sensitivity, 83.3%; or 100% during time worn). Positive predictive value for AF episodes was 39.9%. Conclusions: An AFSW is highly sensitive for detection of AF and assessment of AF duration in an ambulatory population when compared with an ICM. Such devices may represent an inexpensive, noninvasive approach to long-term AF surveillance and management.

IMPORTANCEFor patients with chronic kidney disease (CKD), hyperkalemia is common, associated with fatal arrhythmias, and often asymptomatic, while guideline-directed monitoring of serum potassium is underused. A deep-learning model that enables noninvasive hyperkalemia screening from the electrocardiogram (ECG) may improve detection of this life-threatening condition.OBJECTIVE To evaluate the performance of a deep-learning model in detection of hyperkalemia from the ECG in patients with CKD. DESIGN, SETTING, AND PARTICIPANTSA deep convolutional neural network (DNN) was trained using 1 576 581 ECGs from 449 380 patients seen at Mayo Clinic, Rochester, Minnesota, from 1994 to 2017. The DNN was trained using 2 (leads I and II) or 4 (leads I, II, V3, and V5) ECG leads to detect serum potassium levels of 5.5 mEq/L or less (to convert to millimoles per liter, multiply by 1) and was validated using retrospective data from the Mayo Clinic in Minnesota, Florida, and Arizona. The validation included 61 965 patients with stage 3 or greater CKD. Each patient had a serum potassium count drawn within 4 hours after their ECG was recorded. Data were analyzed between April 12, 2018, and June 25, 2018.EXPOSURES Use of a deep-learning model. MAIN OUTCOMES AND MEASURESArea under the receiver operating characteristic curve (AUC) and sensitivity and specificity, with serum potassium level as the reference standard. The model was evaluated at 2 operating points, 1 for equal specificity and sensitivity and another for high (90%) sensitivity. RESULTSOf the total 1 638 546 ECGs, 908 000 (55%) were from men. The prevalence of hyperkalemia in the 3 validation data sets ranged from 2.6% (n = 1282 of 50 099; Minnesota) to 4.8% (n = 287 of 6011; Florida). Using ECG leads I and II, the AUC of the deep-learning model was 0.883 (95% CI, 0.873-0.893) for Minnesota, 0.860 (95% CI, 0.837-0.883) for Florida, and 0.853 (95% CI, 0.830-0.877) for Arizona. Using a 90% sensitivity operating point, the sensitivity was 90.2% (95% CI, 88.4%-91.7%) and specificity was 63.2% (95% CI, 62.7%-63.6%) for Minnesota; the sensitivity was 91.3% (95% CI, 87.4%-94.3%) and specificity was 54.7% (95% CI, 53.4%-56.0%) for Florida; and the sensitivity was 88.9% (95% CI, 84.5%-92.4%) and specificity was 55.0% (95% CI, 53.7%-56.3%) for Arizona.CONCLUSIONS AND RELEVANCE In this study, using only 2 ECG leads, a deep-learning model detected hyperkalemia in patients with renal disease with an AUC of 0.853 to 0.883. The application of artificial intelligence to the ECG may enable screening for hyperkalemia. Prospective studies are warranted.

The SHM is accurate in measuring QTc interval in sinus rhythm when compared to 12-lead ECG in healthy volunteers. For patients receiving QT prolonging antiarrhythmics, SHM is capable of detecting QTc prolongation, and lead-I of the SHM is most accurate in measuring the QTc if < 500 milliseconds.

Peak aortic blood acceleration is recognized to be a sensitive index of global left ventricular performance. In the present study peak acceleration was assessed noninvasively in patients with a continuous-wave Doppler velocity meter. Peak aortic blood velocity and peak blood acceleration were measured by placing the ultrasonic transducer at the suprastemal notch. Measurements were obtained in 36 patients undergoing diagnostic cardiac catheterization. Peak velocity and acceleration were measured at rest just before left ventriculography. In patients with ejection fractions greater than 60%, peak acceleration was 19 5 m/sec/sec. In patients with ejection fractions of 41% to 60%, peak acceleration was lower, at 12 2 m/sec/sec (p < .001). In patients with ejection fractions of 40% or less, peak acceleration (8 ± 2 mlsec/sec) was markedly lower than in patients with ejection fractions greater than 60% (p <. 001). Peak acceleration showed a good linear correlation with ejection fraction (r = .90), and a better power fit (r = .93